What problem does it solve? Integrating clinical lab report parsing is error-prone: uploads return upload IDs instead of session IDs, webhooks redeliver and reprocess with confusing idempotency semantics, unmatched biomarkers arrive with null keys, and reference ranges vary by patient demographics. This Skill encodes the correct handling for each of these pitfalls so agents generate working Terra Lab Reports integrations. ## Core Features & Use Cases - Async Lifecycle Guidance: Covers the upload/standardize/deliver pipeline, the lab_report.completed / lab_report.failed event envelope, and polling fallbacks with rate-limit-safe intervals. - Layered Data Model Reference: Documents the source / biomarker / measurement / interpretation / reference_ranges result structure, UCUM unit codes, LOINC mappings, and open enum handling. - Best-Practice Rules: Seven incorrect/correct code rules covering webhook deduplication on event_id, null biomarker.key fallback, demographic range filtering, snowflake ID string storage, and UTF-8 parsing. - Use Case: Build a webhook handler that dedupes redeliveries on event_id, stores snowflake session IDs as strings, keeps unmatched biomarkers with a source.name fallback, and filters reference ranges by patient sex and age. ## Quick Start Ask your agent to write a webhook handler for Terra lab report results that deduplicates on event_id and stores the parsed biomarker results in a database.
